Risky sexual behavior (RSB) is the leading cause of HIV worldwide. Evidence suggests that drug use is related to engagement in RSB, with this relationship differing across drug of choice. Indeed, several researchers have documented that RSB is significantlymore frequent in heroin users who use crack/cocaine compared to heroin users who do use crack/cocaine. These studies have been interpreted to suggest that crack/cocaine use is a risk-factor for RSB. One limitation, however, is that heroin users also using crack/cocaine are using two drugs whereas heroin users not using crack/cocaine are using only one drug. Thus, greater levels of crack/cocaine use in these studies also could be conceptualized as the additive effects of another drug in general, and not necessarily the specific effects of crack/cocaine. Indeed, a true test of the role of drug choice requires independent groups (e.g., individuals dependent on crack/cocaine and not heroin vs. individuals dependent on heroin and not crack/cocaine vs. individuals dependent on both crack/cocaine and heroin) to more clearly isolate resulting differences. As a second limitation of this line of research, few studies have provided a clearly developed theoretical framework to explain why increased levels of RSB are related to crack/cocaine use. The current application will compare level of RSB across the following three groups of individuals beginning residential drug use treatment: a) 100 individuals dependent on crack/cocaine and using heroin no more than once monthly (covering a period 6 months prior to treatment); and b) 100 individuals dependent on heroin and using crack/cocaineno more than once monthly (covering a period 6 months prior to treatment); and c) 100 individuals dependent on both crack/cocaine and heroin (covering a period 6 months prior to treatment). Further, the current study will include an examination of the moderating and mediating effects of theory-driven combinations of variables taken from Social Action Theory and impulsivity across several domains of this construct (e.g., delay discounting, behavioral inhibition, impulsive risk taking) to begin to develop a larger framework to understand drug choice and RSB. The results will set the stage for future longitudinal research further assessing issues of causality and will begin to provide useful information regarding the prevention of HIV infection through better understanding risk factors for RSB.